jb-docs-scraper
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- Unverifiable Dependencies (MEDIUM): The skill dynamically installs the
crawl4aipackage and Playwright binaries from external sources at runtime, which are not part of the trusted source whitelist. - Evidence:
SKILL.mdandreferences/scrape_docs.pyrecommend runninguv run --with crawl4aiandplaywright install. - Indirect Prompt Injection (LOW): The skill ingests untrusted data from the web and stores it in a format intended for AI context without sufficient protection.
- Ingestion points:
references/scrape_docs.pyusingAsyncWebCrawler.arun()to fetch content from user-provided URLs. - Boundary markers: Absent. The script writes the scraped markdown directly to files without adding delimiters or warning instructions for the LLM.
- Capability inventory: The script performs file-system writes (
Path.write_text) and network requests. The output is explicitly described as 'AI context', creating a direct path for ingested content to influence the agent's future actions. - Sanitization: Absent. Web content is converted to markdown and saved without filtering for potential malicious instructions targeted at an LLM.
Audit Metadata